Skip to content

Instance Selector

eksctl supports specifying multiple instance types for managed and self-managed nodegroups, but with over 270 EC2 instance types, users have to spend time figuring out which instance types would be well suited for their nodegroup. It's even harder when using Spot instances because you need to choose a set of instances that works together well with the Cluster Autoscaler.

eksctl now integrates with the EC2 instance selector, which addresses this problem by generating a list of instance types based on resource criteria such as vCPUs, memory, etc. When the instance selector criteria is passed, eksctl creates a nodegroup with the instance types set to the instance types matching the supplied criteria.

Create cluster and nodegroups

To create a cluster with a single nodegroup that uses instance types matched by the instance selector resource criteria passed to eksctl, run

$ eksctl create cluster --instance-selector-vcpus=2 --instance-selector-memory=4

This will create a cluster and nodegroup with the instancesDistribution.instanceTypes field set to [c5.large, c5a.large, c5ad.large, c5d.large, t2.medium, t3.medium, t3a.medium] (the set of instance types returned may change).

For managed nodegroups, the instanceTypes field will be set:

$ eksctl create cluster --managed --instance-selector-vcpus=2 --instance-selector-memory=4

The instance selector criteria can also be specified in ClusterConfig:

# instance-selector-cluster.yaml
---
apiVersion: eksctl.io/v1alpha5
kind: ClusterConfig

metadata:
  name: cluster
  region: us-west-2

nodeGroups:
- name: ng
  instanceSelector:
    vCPUs: 2
    memory: "4" # 4 GiB, unit defaults to GiB

managedNodeGroups:
- name: mng
  instanceSelector:
    vCPUs: 2
    memory: 2GiB #
    cpuArchitecture: x86_64 # default value
$ eksctl create cluster -f instance-selector-cluster.yaml

The following instance selector CLI options are supported by eksctl create cluster and eksctl create nodegroup:

--instance-selector-vcpus, --instance-selector-memory, --instance-selector-gpus and instance-selector-cpu-architecture

An example file can be found here.

Dry Run

The dry-run feature allows you to inspect and change the instances matched by the instance selector before proceeding to creating a nodegroup.

$ eksctl create cluster --name development --managed --instance-selector-vcpus=2 --instance-selector-memory=4 --dry-run

apiVersion: eksctl.io/v1alpha5
kind: ClusterConfig
# ...
managedNodeGroups:
- amiFamily: AmazonLinux2
  instanceSelector:
    memory: "4"
    vCPUs: 2
  instanceTypes:
  - c5.large
  - c5a.large
  - c5ad.large
  - c5d.large
  - t2.medium
  - t3.medium
  - t3a.medium
...
# other config

The generated ClusterConfig can then be passed to eksctl create cluster:

$ eksctl create cluster -f generated-cluster.yaml

The instanceSelector field representing the CLI options will also be added to the ClusterConfig file for visibility and documentation purposes. When --dry-run is omitted, this field will be ignored and the instanceTypes field will be used, otherwise any changes to instanceTypes would get overridden by eksctl.

When a ClusterConfig file is passed with --dry-run, eksctl will output a ClusterConfig file containing the same set of nodegroups after expanding each nodegroup's instance selector resource criteria.

# instance-selector-cluster.yaml
---
apiVersion: eksctl.io/v1alpha5
kind: ClusterConfig

metadata:
  name: cluster
  region: us-west-2

nodeGroups:
- name: ng
  instanceSelector:
    vCPUs: 2
    memory: 4 # 4 GiB, unit defaults to GiB

managedNodeGroups:
- name: mng
  instanceSelector:
    vCPUs: 2
    memory: 2GiB #
    cpuArchitecture: x86_64 # default value
$ eksctl create cluster -f instance-selector-cluster.yaml --dry-run

apiVersion: eksctl.io/v1alpha5
kind: ClusterConfig
# ...
managedNodeGroups:
- amiFamily: AmazonLinux2
  # ...
  instanceSelector:
    cpuArchitecture: x86_64
    memory: 2GiB
    vCPUs: 2
  instanceTypes:
  - t3.small
  - t3a.small
nodeGroups:
- amiFamily: AmazonLinux2
  # ...
  instanceSelector:
    memory: "4"
    vCPUs: 2
  instanceType: mixed
  instancesDistribution:
    capacityRebalance: false
    instanceTypes:
    - c5.large
    - c5a.large
    - c5ad.large
    - c5d.large
    - t2.medium
    - t3.medium
    - t3a.medium
# ...